Repository hosted by TU Delft Library

Home · Contact · About · Disclaimer ·

A multi-component data assimilation experiment directed to sulphur dioxide and sulphate over Europe

Publication files not online:

Author: Barbu, A.L. · Segers, A.J. · Schaap, M. · Heemink, A.W. · Builtjes, P.J.H.
Institution: TNO Bouw en Ondergrond
Source:Atmospheric Environment, 9, 43, 1622-1631
Identifier: 241423
Keywords: Environment · Data assimilation · Ensemble Kalman filter · LOTOS-EUROS model · Sulphate · Sulphur dioxide · Air filters · Applications · Control theory · Experiments · Kalman filters · Nonmetals · Permanent magnets · Sulfur · Sulfur determination · Uncertainty analysis · Data assimilation · Ensemble Kalman filter · LOTOS-EUROS model · Sulphate · Sulphur dioxide · Sulfur dioxide · atmospheric modeling · climate effect · data assimilation · health impact · Kalman filter · particulate matter · pollutant transport · pollution monitoring · sulfate · sulfur dioxide · Eurasia · Europe


Fine particulate matter (PM) is relevant for human health and its components are associated with climate effects. The performance of chemistry transport models for PM, its components and precursor gases is relatively poor. The use of these models to assess the state of the atmosphere can be strengthened using data assimilation. This study focuses on simultaneous assimilation of sulphate and its precursor gas sulphur dioxide into the regional chemistry transport model LOTOS-EUROS using an ensemble Kalman filter. The process of going from a single component setup for SO2 or SO4 to an experiment in which both components are assimilated simultaneously is illustrated. In these experiments, solely emissions, or a combination of emissions and the conversion rates between SO2 and SO4 were considered uncertain. In general, the use of sequential data assimilation for the estimation of the sulphur dioxide and sulphate distribution over Europe is shown to be beneficial. However, the single component experiments gave contradicting results in direction in which the emissions are adjusted by the filter showing the limitations of such applications. The estimates of the pollutant concentrations in a multi-component assimilation have found to be more realistic. We discuss the behavior of the assimilation system for this application. The model uncertainty definition is shown to be a critical parameter. The increased complexity associated with the simultaneous assimilation of strongly related species requires a very careful specification of the experiment, which will be the main challenge in the future data assimilation applications. © 2008 Elsevier Ltd. All rights reserved.